J Korean Radiol Soc 2007;56:491-495

─ 491 ─

MR Imaging of Endometrial Cancer that Occurs AfterRadiation Therapy for Cervix Cancer1

Youn Jeong Kim, M.D., Yong Yeon Jeong, M.D.2, Nam Yeol Lim, M.D.2, Seok Wan Ko, M.D.3, Bo Hyun Kim, M.D.4

1Department of Diagnostic Radiology, Inha University Hospital, Korea 2Department of Diagnostic Radiology, Chonnam National University Medical School, Chonnam National University Hwasun Hospital3Department of Diagnostic Radiology, Kwangju Christian Hospital, Korea4Department of Diagnostic Radiology, Mayo Clinic College of Medicine, Rochester, Minnesota, USA.Received December 19, 2006 ; Accepted February 28, 2007Address reprint requests to : Yong Yeon Jeong, M.D., Department of Radiology, Chonnam National University Medical School, Chonnam NationalUniversity Hwasun Hospital, 160, Ilshim-li Hwasun-eup, Hwasun-gun, Jeollanam-Do, Korea. Tel. 82-61-379-7102 Fax. 82-61-379-7133 E-mail: yjeong@chonnam.ac.kr

Purpose: We wanted to describe the MR imaging findings of endometrial cancer in pa-tients with a history of prior radiation therapy for cervical cancer (ECRT) and we com-pare them to the MR imaging findings of patients with spontaneously occurring en-dometrial cancer (SEC).Materials and Methods: Twenty-two patients with endometrial cancer that was diag-nosed by operation or endometrial biopsy were included in the study. The patientswere divided into two groups according to the presence of past RT for cervical cancer:ECRT (n = 4) and SEC (n = 18). The MR images were retrospectively analyzed by con-sensus of two experienced radiologists. The MR imaging findings were analyzed bythe size, shape and signal intensity of the mass, distension of the uterine cavity, thepresence of cervical stenosis and the nature of the fluid collection. Results: For the mass shape, all the ECRT lesions were polypoid masses. However, theSEC patients had 5 polypoid masses and 13 wall thickenings. The maximal diameter,signal intensity and enhancement pattern of the masses were not different betweenthe ECRT and SEC patients. The width of the endometrial cavity varied between 3.9cm in the ECRT patients and 0.4 cm in the SEC patients (p =0.002). All the ECRT pa-tients had cervical stenosis. However, none of the SEC patients had cervical stenosis. Conclusion: MR imaging of ECRT patients demonstrated prominent distension oftheir uterine cavity and cervical stenosis, which may be the result of radiation fibrosisin the uterus.

Index words : Uterine neoplasms, MR Uterus, endometriumRadiationsInjurious effectsComplications of therapeutic radiology

Endometrial cancer is the seventh most common ma-lignant disorder worldwide, but its incidence varies be-tween different geographic regions (1). Clinically, pa-tients who have endometrial cancer generally presentwith abnormal uterine bleeding. The findings, whenperforming endocervical curettage and suction endome-trial biopsy, confirm the diagnosis of endometrial cancerand provide the tumor grade and histologic type (2, 3).MR imaging has only a small role in detecting endome-trial cancer. However, MR imaging is considered themost accurate imaging technique for the preoperativeassessment of endometrial cancer due to its excellentsoft-tissue contrast (4-7).

Radiation therapy (RT) is the standard treatment formost patients with stages IIB-IVA cervical cancer (8).The patients who have survived after RT for cervicalcancer have frequently developed secondary malignan-cies such as uterine sarcoma, bladder cancer, vaginalcancer and cancer of the overlying integument becauseRT treatment is relatively successful and many patientssurvive long enough to be at risk for the late complica-tions from radiotherapy. The incidence of endometrialcancer, as a second malignancy following radiation, isvery low. The endometrial carcinomas arising after RTfor treating cervical cancers are poorly differentiatedand they are usually diagnosed at an advanced stage (9-11). We hypothesized that the MR imaging findings ofendometrial cancer in patients with a history of prior ra-diation therapy for cervical cancer (ECRT) can be differ-entiated from the MR imaging findings of spontaneouslyoccurring endometrial cancer (SEC). SEC patients usual-ly undergo MRI and they have a known histological di-agnosis via endometrial biopsy. In contrast, ECRT pa-tients, because of their cervical stenosis, cannot undergoendometrial biopsy and they are being imaged whencancer is suspected. Thus, for ECRT patients, the radiol-ogist must make the initial diagnosis of cancer. The pur-pose of this study was to describe the MR imaging find-ings of ECRT patients and we compared these findingswith the MR findings of SEC patients.

Materials and Methods

The records in the radiology-pathology databases fromOctober 2002 to September 2004 at two institutions in-cluded four ECRT patients. From October 2002 toSeptember 2004, 18 consecutive patients with SEC with-out RT were retrospectively evaluated at one institution.All the patients were referred to the radiology depart-

ment for MR imaging. Their mean age was 55.4±12.8years. All the patients had their diagnosis confirmed byendometrial biopsy (n=3) or operation (n=19). The pa-tients were divided into two groups according to thepast performance of RT. The mean age of the ECRT pa-tients was 53.5±4.2 and the mean age of the SEC pa-tients was 55.8±14.0. The mean latency period of theECRT was 10.3 years. The stage was determined by ret-rospectively applying the FIGO staging criteria via theoperative and pathologic findings at surgery for theECRT patients and via the imaging and biopsy patholog-ic findings in the other SEC patients (12).

All the patients underwent MR imaging with a 1.5TMR scanner (GE Signa Horizon; GE Healthcare,Milwaukee, WI, U.S.A.) and with using a body coil. TheT1-weighted spine-echo images with a repetition time of500 ms and an echo time of 8 ms were obtained in theaxial plane. The T2-weighted spine-echo images with arepetition time of 3,500-3,800 ms and an echo time of80-90 ms were obtained in the axial and sagittal planes.The Imaging parameters were a 5 mm section thickness,a 1 mm intersection gap and a 512×224 matrix. The con-trast-enhanced, fat-suppressed, gradient echo T1-weight-ed images (TR/TE = 120/1.7 ms, Flip angle = 90°, 5 mmsection thickness) were obtained after the administrationof 0.1 mmol gadolinium per kilogram of body weight.

The shape of the mass, the maximal diameter of themass, the signal intensity of the mass and the enhance-ment pattern of the mass, the width of the endometrialcavity and the signal intensity of the fluid on the MR im-ages were all analyzed by two experienced radiologistsworking in consensus. The shape of mass was classifiedas an irregular polypoid mass and/or wall thickening.When a mass was pedunculated, which resulted in anacute angle between the mass and the uterine wall, thenthe shape of the mass was considered as a polypoidmass. The maximal diameter of the mass was classifiedaccording to the thickest diameter of the thick walled le-sion and the maximal size of the polypoid mass. The sig-nal intensity of the mass was classified into hyperin-tense, hypointense or isointense, relative to the signal in-tensity of the myometrium.

The size of the mass and the width of the uterine bodybetween the ECRTs and SECs were analyzed with usingthe Mann-Whitney U test.

Results

The distribution of the tumor stages for the ECRT was

Youn Jeong Kim, et al : MR Imaging of Endometrial Cancer that Occurs After Radiation Therapy for Cervix Cancer

─ 492 ─

stage I: 1, stage II: 1, stage III: 2 and stage IV: 1. All theSEC patients were stage I. The histologic types of ECRTwere as follows; subtypes of papillary serous carcinoma:2, clear cell carcinoma: 1 and endometroid carcinoma:1. In contrast, all the SEC patients had endometroid car-cinoma.

Regarding the mass shape, all lesions of the ECRT pa-tients were polypoid masses (Fig. 1). However, the SEC

lesions showed five polypoid masses and 13 wall thick-enings (Fig. 2). The maximal diameter of the masses var-ied between 1 cm and 2.6 cm (mean±SD: 1.53±0.75cm) for the ECRT patients and between 0.5 cm and 3.5cm (mean±SD: 2.25±1.07 cm) for the SEC patients (p= 0.157).

For the SEC patients, all the lesions of the 22 patientsdemonstrated iso-signal intensity lesion on the T1-

J Korean Radiol Soc 2007;56:491-495

─ 493 ─

A BFig. 1. A 48-year-old woman with endometrial cancer 14 years after radiation therapy. A. The axial T1-weighted MR image shows multiple, polypoid isointense masses (arrows) within the distended endometrial cavity.The signal intensity of the fluid within the endometrial cavity is hyperintense, which is suggestive of hematoma. B. The sagittal T2-weighted MR image reveals polypoid, hyperintense masses (arrows) in the markedly distended endometrial cavi-ty.

A BFig. 2. A 48-year-old woman with spontaneous endometrial cancer. A. The sagittal T2-weighted image shows the hyperintense thickened lesion of the endometrium (arrows).B. The contrast-enhanced gradient echo sagittal T1-weighted image demonstrates the homogenous enhancing endometrial cancer(arrows) with an intact junction zone.

weighted images and high signal intensity lesion on theT2-weighted images. For the ECRT patients, three offour patients (75%) showed heterogeneous enhance-ment and one patient showed homogeneous enhance-ment on the contrast enhanced T1-weighted images.However, for the SEC patients, 13 of 18 patients (72.2%)showed heterogeneous enhancement.

The width of the endometrial cavity varied between1.6 cm and 5.2 cm (mean±SD: 3.9±1.7 cm) in theECRT patients, whereas for the SEC patients, the widthof the endometrial cavity varied between 0.1 cm and 2.7cm (mean±SD: 0.6±0.7 cm). The width of the endome-trial cavity in the ECRT patients was wider than that ofthe SEC patients (p=0.002) (Fig. 1). In all patients, thesignal intensity of the fluid in the endometrial cavity ap-peared cystic as hypointensity on the T1-weighed im-ages and as hyperintensity on the T2-weighted images,except for 2 ECRT patients. These 2 patients hadhematoma in the endometrial cavity.

All ECRT patients had cervical stenosis and this wasestablished via pelvic exams. However, there was nocervical stenosis in all the SEC patients.

Discussion

Endometrial tissue can persist after administering RTfor cervical carcinoma, and so the risk of developing en-dometrial carcinoma must be considered. ECRT maydevelop from the remnant endometrial tissue; however,theses cases are very rare. Most ECRT patients areasymptomatic, but the advanced cases complain of low-er abdominal distension or pain. The MRI findings ofECRT are a polypoid mass with prominent distension ofthe endometrial cavity or a distended endometrial cavi-ty with a large tumor filling it. The prognosis of ECRT ispoorer than that for SEC.

Several studies have reported that aggressive adeno-carcinomas developed after administering RT for cervi-cal carcinoma, implicating radiation as a carcinogenicfactor for the development of this aggressive histologicalsubtype (10, 13). In our study, 75% of the patients (3/4)had aggressive subtypes of adenocarcinoma such aspapillary serous carcinoma or clear cell carcinoma.Papillary serous carcinoma and clear cell carcinoma aresignificantly more common in black women. Patientswith papillary serous carcinoma and clear cell carcino-ma are more likely to have pelvic and periaortic tumorinvolvement (14).

The significance of ionizing radiation as a risk factor

for the development of endometrial cancer is currentlyunresolved. However, given the age of diagnosis, the la-tent period and the preponderance of high-risk histolog-ic types, radiation is likely to be the cause. Therefore, itis important to continue careful annular surveillanceeven if the patient has been free of disease for manyyears, as the mean latency period for developing en-dometrial cancer is 10-14 years (9, 13). In our study,endometrial cancers were diagnosed at a mean of 10.3years after administering radiation for cervical cancer,supporting that this should be sufficient time to excludethe presence of this tumor at the time of irradiation (9-11).

Radiation-induced cervical stenosis may prevent thepatient from displaying early symptoms and this compli-cates efforts to obtain a tissue sample for diagnosis.Unlike the SEC patients, whose first symptoms are typi-cally vaginal bleeding, the patients with ECRT usuallypresented with the signs and symptoms of an enlargeuterus and pelvic pain, indicating relatively advanceddisease (13). In our study, three ECRT patients (75%)had their disease detected at advanced stages (stages II-IV), whereas all the cases of SEC were stage I.

The MR images of endometrial carcinoma haveshown various abnormal endometrial findings. The en-dometrium may be thickened focally or diffusely, andthis seen as being irregular in thickness and configura-tion, or as widened via polypoid tumor (5, 7). Polypoidmasses were revealed in all ECRT patients (100%),whereas polypoid masses were seen in five SEC patients(27.8%) in our study. It seems to us that a polypoid masswas seen only in the ECRT patients who have hydrome-tra. On the other hand, the endometrial thickening wasmore commonly seen in SEC patients whose cavitieswere collapsed.

The signal intensity of the tumor shows various pat-terns on the T1-weighted and T2-weighted images (7). Inour study, although the signal intensity and enhance-ment pattern on the MR imaging were not different be-tween the ECRT and SEC patients, prominent wideningof the endometrial cavity was revealed in the ECRT pa-tients compared with the SEC patients. The cause of thismay be radiation fibrosis with cervical stenosis.

The limitation of this study is the small number ofECRT patients. Because the development of ECRT isquite rare, only 4 patients belonging to this group wereincluded in this study. A future study is needed to de-scribe the MR imaging in a large number of ECRT pa-tients.

Youn Jeong Kim, et al : MR Imaging of Endometrial Cancer that Occurs After Radiation Therapy for Cervix Cancer

─ 494 ─

In summary, it is important to recognize that ECRTpatients may show a high-risk histological subtype and ahistory of pelvic radiation for treating cervical cancer.The MR imaging findings of ECRT patients showed apolypoid mass with prominent distension of the en-dometrial cavity, as compared with SEC patients.

References

1. Hoffman M, Roberts WS, Cavanagh D. Second pelvic malignan-cies following radiation therapy for cervical cancer. Obstet GynecolSurv 1985;40:611-617

2. Ben-Shachar I, Pavelka J, Cohn DE, Copeland LJ, Ramirez N,Manolitsas T, et al. Surgical staging for patients presenting withgrade 1 endometrial carcinoma. Obstet Gyencol 2005;105:487-493

3. Daniel AG, Peters WA 3rd. Accuracy of office and operating roomcurettage in the grading of endometrial carcinoma. Obstet Gynecol1988;71:612-614

4. Yamashita Y, Harada M, Sawada T, Takahashi M, Miyazaki K,Okamura H. Normal uterus and FIGO stage I endometrial carcino-ma: dynamic gadolinium-enhanced MR imaging. Radiology 1993;186:495-501

5. Sironi S, Colombo E, Villa G, Taccagni G, Belloni C, Garancini P,et al. Myometrial invasion by endometrial carcinoma: assessmentwith plain and gadolinium-enhanced MR imaging. Radiology 1992;185:207-212

6. Ito K, Matsumoto T, Nakada T, Nakanishi T, Fujita N, YamashitaH. Assessing myometrial invasion by endometrial carcinoma withdynamic MRI. J Comput Assist Tomogr 1994;18:77-86

7. Manfredi R, Gui B, Maresca G, Fanfani F, Bonomo L. Endometrialcancer: magnetic resonance imaging. Abdom Imaging 2005;30:626-636

8. Morris M, Eifel PJ, Lu J, Grigsby PW, Levenback C, Stevens RE, etal. Pelvic radiation with concurrent chemotherapy compared withpelvic and para-aortic radiation for high-risk cervical cancer. NEngl J Med 1999;340:1137-1143

9. Fehr PE, Prem KA. Malignancy of the uterine corpus following ir-radiation therapy for squamous cell carcinoma of the cervix, Am JObstet Gynecol 1974;119:685-692

10. Gallion HH, van Nagell JR Jr, Donaldson ES, Powell DE.Endometrial cancer following radiation therapy for cervical carci-noma. Gynecol Oncol 1987;27:76-83

11. Rodriguez J, Hart WR. Endometrial cancers occurring 10 or moreyears after pelvic irradiation for carcinoma. Int J Gyencol Pathol1982;1:135-144

12. Creasman WT. Malignant tumors of the uterine corpus. In: Rock JA,Jones HW, ed. Te Lind’s Operative gynecology, 9th ed. Philadelphia:Lippincott Williams & Wilkins, 2003;1445-1486

13. Pothuri B, Ramondetta L, Martino M, Alektiar K, Eifel PJ, DeaversMT, et al. Development of endometrial cancer after radiation treat-ment for cervical carcinoma. Obstet Gynecol 2003; 101:941-945

14. Matthews RP, Hutchinson-Colas J, Maiman M, Fruchter RG,Gates EJ, Gibbon D, et al. Papilary serous and clear cell type leadto poor prognosis of endometrial carcinoma in black women.Gynecol Oncol 1997;65:206-212

J Korean Radiol Soc 2007;56:491-495

─ 495 ─

대한영상의학회지 2007;56:491-495

자궁 경부암으로 방사선 치료 후 발생한자궁 내막암종의 자기공명영상소견1

1인하대병원 방사선과, 2전남대 화순병원 방사선과, 3광주기독병원 방사선과, 4Myo대학병원

김윤정·정용연2·임남열2·고석완3·김보현4

목적: 자궁 경부암으로 방사선치료를 받았던 환자들에서 발생한 자궁 내막암종의 자기공명영상소견을 기술하고 자

발적으로 발생하는 자궁 내막암종과의 차이를 알아보고자 한다.

대상과 방법: 수술 혹은 자궁내막 조직검사에 의해 발견된 22명의 자궁 내막암 환자를 대상으로 하였다. 환자들을

자궁 경부암으로 방사선치료를 받았던 과거력의 유무에 따라 나누었을 때 각각 4명(방사선 치료 후), 18명(자연발

생)이었다. 자기공명영상 소견은 종괴의 크기, 모양, 신호강도와 자궁강 확장, 자궁협부협착, 자궁내 저류된 액체성

상을 분석하였다.

결과: 방사선 치료 후는 모든 종괴 모양은 폴립 양인 데 비해 자연발생은 5명만 폴립 양이고 나머지 13명은 두꺼

운 내벽으로 보였다. 최대 크기, 신호강도, 종괴의 조영증강은 두 그룹 간에 차이가 없었다. 자궁 강이 확장된 너비

는 방사선 치료 후에서는 3.9 cm이고 자연발생에서는 0.4 cm으로 통계학적으로 유의하였다. 모든 방사선 치료 후

의 환자들은 자궁협부가 협착되었고 반면, 자연발생의 환자들은 자궁협부협착을 볼 수 없었다.

결론: 자궁 경부암으로 방사선치료를 받았던 환자에서 발생한 자궁 내막암종은 자발적으로 생긴 자궁 내막암종에

비해 자궁강 확장과 자궁협부협착의 자기공명영상소견을 보였다. 자궁협부협착은 방사선 치료에 따른 섬유화 변화

에 의한 소견일 것이다.